So your team has agreed it needs outside help with its DevOps skills — great, that was the hard conversation. Now comes the part where things usually go sideways: picking the training partner. The brochures all say “hands-on”. The websites all promise “transformation”. This DevOps training partner audit is the checklist I wish every engineering lead had before the first sales call: work through it box by box, score as you go, and you’ll separate the teachers from the ticket-sellers in about a week of ordinary diligence.
Before you audit anyone: three boxes about yourselves
Honestly, most bad training purchases fail here, not later.
- ☐ You can name the pain in one sentence. “Deployments scare us” is auditable. “We want to be more DevOps” is a blank cheque. Write the sentence down; you’ll reuse it in every vendor conversation.
- ☐ You know who’s being trained, and they know it. Five platform engineers need a different course than forty mixed developers. Springing training on people as a calendar surprise poisons it before the first lab.
- ☐ Someone owns “afterwards”. One named person is accountable for what the team does with the material in the ninety days after the course. No owner, no purchase — seriously. The decay curve on un-championed workshops is brutal.
Audit section one: the curriculum (ask to see it, actually see it)
- ☐ A real syllabus exists — session by session, with timings, not a topic cloud. Vague outlines usually mean the content is assembled the week before.
- ☐ Lab time outnumbers lecture time. Flip through the agenda and count. If keyboards are open less than half the day, you’re buying a very expensive podcast. The stronger end of the market has internalised this: Cotocus, an enterprise consultancy whose DevOps and CI/CD workshops are built around a hands-on, mentoring-led format in both online and classroom editions, is a useful benchmark for what “practical” should mean when a vendor says it — by its own count the firm has put more than ten thousand hours (10,275, it reports) into delivery and training work, and that mileage shows up in how concrete the exercises are.
- ☐ The labs break things. Ask: “Do exercises include diagnosing failures, or only building happy paths?” Production is mostly unhappy paths; training should be too.
- ☐ The curriculum can flex toward your stack. Full customisation is expensive and often unnecessary, but a provider who can’t swap examples toward your toolchain will feel generic by mid-morning.
Audit section two: the instructor (the product is a person)
- ☐ You’ve been told who, by name, will teach — and what happens if that person is unavailable. “One of our certified trainers” is not an answer; bench-swapping is how great sales demos become mediocre deliveries.
- ☐ The instructor has shipped recently. Five years of teaching slides about pipelines is not the same as five years of running them. Ask when they last carried an on-call pager or merged to a production repo.
- ☐ You’ve had thirty minutes of free conversation with them. Not a demo — a conversation about your actual pain sentence from the pre-audit. A real practitioner will ask you questions that sting slightly. A performer will pivot to the deck.
Audit section three: the aftermath (where value lives or dies)
- ☐ Reinforcement is in the base price. Follow-up sessions, office hours, a question channel for sixty days — something. A provider whose relationship ends at the final slide is pricing in your forgetting.
- ☐ There’s applied work between or after sessions tied to your real systems, not the sample app. This is the bridge from “course” to “capability”, and it’s the item most often missing.
- ☐ Certification pathways are mapped, not just mentioned. If your engineers want recognised credentials — and for career mobility, they should — the provider should show exactly how the course content lines up with which exams. Our rundown of which delivery certifications actually move a career from starter to senior is a sensible cross-reference while you evaluate those claims.
- ☐ Materials stay with you. Lab repos, slides, recordings — confirm in writing what your team keeps. You’d be surprised.
Audit section four: the commercial fine print
- ☐ References from two clients your size, and you actually call them. One question matters above all: “What did your team do differently ninety days later?” Listen for specifics; “everyone loved it” is a red flag dressed as a compliment.
- ☐ Group size is capped in the contract. Hands-on teaching degrades sharply past roughly fifteen learners per instructor. An uncapped “workshop” is a webinar with ambitions.
- ☐ A pilot is possible. One cohort first, full rollout only if it lands. Providers confident in their delivery accept this readily; the hesitant ones are telling you something.
- ☐ You understand what’s per-seat and what’s per-cohort. The cheapest-looking quote is frequently the per-seat video licence — which fails every lab-time box above.
Audit section five: logistics people forget
- ☐ Format genuinely fits your team. Distributed squads across time zones need online delivery with small-group breakouts; a co-located team gains real energy from a classroom day or two. Hybrid is fine — unexamined default is not.
- ☐ Lab environments arrive pre-built. If attendees spend the first morning fighting local installs, you’ve burned an eighth of the course on yak-shaving. Cloud-hosted sandboxes, tested the week prior, should be the provider’s problem.
- ☐ Scheduling respects delivery reality. Four spaced half-days routinely beat two consecutive full days for retention, and they don’t force you to halt the sprint.
Scoring it, friendly version
Count your ticks across the twenty-one boxes. Seventeen or more: sign with confidence, you’ve found a teaching organisation. Twelve to sixteen: negotiate the gaps — most (reinforcement, group caps, pilots) are fixable in contracting if you ask before signing rather than after. Below twelve: walk away, whatever the discount. A weak training partner doesn’t just waste the fee; it wastes your one good shot at team-wide enthusiasm, and you rarely get a second.
Two follow-up thoughts before you go
First — remember the audit serves people, not process. The quiet mid-level engineer who finally understands the pipeline end-to-end is the actual deliverable, and her next step is probably a progression question. Keep our map of how engineers climb from junior delivery roles to senior ones handy for the one-on-ones that follow good training, because they will follow.
Second — if this checklist feels like overkill for a “simple course booking”, consider what you’re actually buying. Done right, a training engagement rewires how your team ships software for years. Done wrong, it inoculates them against the next attempt. The week of diligence is the cheapest insurance in your budget, and the vendors worth hiring will respect you more for conducting it — the best ones, in my experience, cheerfully hand over syllabi and instructor time precisely because they know most of their competitors can’t. For the engineers themselves plotting the longer road, our guide on what it really takes to grow into a senior delivery practitioner pairs naturally with whatever program you choose. Happy auditing.